Correlation Comparison Test Results

To calculate SMEs' GIC scores, the Knowledge Graph and corresponding network activation metrics have been estimated for each firm. To corroborate the choice of Global Efficiency as GIC measure, Pearson correlations have been calculated between each network metric and the study variables EC, INC, and EP. Then, pairwise comparisons have been performed to highlight significant differences between the correlations of Global Efficiency with the study variables, and those of the other activation metrics (e.g., Density, Global Clustering Coefficient, Small-Worldness Index) with study variables. As reported in Table I, the Correlation Comparison test results confirmed that GE is the network metric showing the strongest correlation with EC, INC, and EP, compared with those of D, CC, and SWI. Moreover, such differences between the correlation coefficients of GE and those of the other network metrics are found to be significant in the majority of the comparisons performed. Consequently, the Global Efficiency has been confirmed as the network activation metric which better performs in association with the study variables. Thus, GE has been adopted as the measure of organizational GIC and used for subsequent analysis.

Table I: Correlations Comparison Test Results

Variable Metric \(r_{jk}\) \(r_{jh}\) \(r_{jk} - r_{jh}\) \(z\) \(p\)
Note. The correlation coefficient \(r_{jk}\) reports the Pearson correlation between GE and each of the study variables (EC, INC, and EP). The \(r_{jh}\) coefficient reports Pearson correlation of the other KG activation metrics (D, CC and SWI) with each of the study variables; the Pearson and Filon’s \(z\) test statistics and corresponding p-values resulting from the comparison \(r_{jk} − r_{jh}\).